Page:Special 301 Report 1994.pdf/6



The USG initiated a Super 301 investigation into the supercomputer procurement practices of the Government of Japan in 1989 which resulted in the 1990 U.S. - Japan Supercomputer Agreement. Positive results under the Agreement, however, have been slow in coming. U.S. machines won only three out of 11 awards made during the first three years of the agreement, all in politically directed, uncontested bids. Further, the USG identified numerous serious problems in Japan's implementation of the Agreement which led USTR to announce, on April 30, 1993, a special review of Japan's compliance with the Agreement under section 306 of the 1974 Trade Act.

Subsequent to the initiation of the review, the Japanese Government procured 15 supercomputers under the regular and supplemental budgets of the 1993 Japan Fiscal Year. Five U.S. supercomputer vendors were awarded six of these contracts. The section 306 review focused heavily on these 15 procurements. While acknowledging that the six U.S. awards represented a positive development in this important sector, the review highlighted several areas of serious concern which demonstrate the need for continued scrutiny of procurements under the Agreement, in order to sustain and build upon this limited progress and achieve a genuinely "fair, open and non-discriminatory" Japanese procurement regime.

U.S. supercomputer vendors dominate the markets outside of Japan, and have a significantly greater share of the Japanese private sector market than of the public sector market. In Europe, U.S. vendors hold an 85 percent share of the public sector market. Despite the demonstrated competitiveness of U.S. machines in all other markets, U.S. supercomputers have never won a head-to-head procurement in the Japanese public market against the same Japanese vector machines they regularly best elsewhere.

This continued lack of head-to-head wins by U.S. vendors in the Japanese public sector is a serious concern that in itself raises questions about Japanese compliance with the Agreement. In addition to this concern, the 306 review surfaced four other areas of ongoing major concern: 1) the notable failure of the Japanese Government to test or evaluate fully systems software, which is essential to the actual performance of a supercomputer and which is well known area of strength among U.S. vendors; 2) concern's about Japan's compliance with the Agreement's provisions on pricing, including the protections against unfairly low bids and the use of market research to establish a reference point for bid prices; 3) results in the Japanese public sector market which do not have a defensible relation to results in other markets, even with results in the Japanese private sector market and 4) anomalies in the conduct of benchmark testing.

Accordingly, the section 306 review will remain underway for an indefinite period of time while we observe additional procurements and consult with the Japanese Government to definitively address these problem areas.